I don’t understand the problem you described whereby “you’ve been investing for ten years but have built up the investment by adding to it each year, you could have the same average annualised return if the early years were excellent and the latter years poor, or vice versa.”

Maybe the XIRR function solves the problem you describe (I think it does).

XIRR

More on XIRR

Anyways, yes I’m doing this for fun.

I picked up a copy of Quicken 2006 for free (stuck on the front of a $6 investment magazine).

ps. Even Quicken doesn’t calculate annualised returns perfectly. It gets very messy when you are adding to or selling part of an existing position - you have to track individual lots. You also can’t just use arithmetic averaging of returns.

I think the easiest method to follow is to just pretend your portfolio is a big ‘black box’, and record all dollar amounts in and out and then iteratively calculate what annual rate of return applied to the model results in the same ending value as your real portfolio. However, unless you have kept a fairly constant amount invested over the total period, this figure is pretty meaningless. eg. If you’ve been investing for ten years but have built up the investment by adding to it each year, you could have the same average annualised return if the early years were excellent and the latter years poor, or vice versa. But althought the average annualised returns are the same the end result is different! To adjust for this you’d have to WEIGHT the return calculation based on the amount you had invested at any time… It all gets way too hard to be of use. You’re better off concentrating on saving more in the early years, and minimising transaction and management costs once you have a reasonable amount invested (say > 5x your annual salary).